Security is an aspect that every organization should give the utmost priority. Ideally, every employee, from the end-user to top-level management are educated when it comes to the impact of network security failure. That said, organizations spend significant capex on securing the network. Despite all the investment on intrusion detection devices, firewalls and access control rules, hackers and their threats continue to succeed—data is stolen, critical services are brought down, and malware manages to sneak into secured networks.
Akamai released their fourth quarter “State of the Internet” report last month which provides valuable insights into, well…obviously, the state of Internet! The security section of the report discusses the top originating country for attack traffic (no points for guessing), the most targeted port, and information about DDoS attacks.
As per the report, the most targeted port for attacks is the good old Telnet port. In fact, Port 23 remains the most targeted port for the 3rd consecutive quarter and attacks against port 23 have increased to 32% from 12% in Q3 2014! This despite the fact that most enterprises I know have shifted to SSH from Telnet to enhance security. The cause of attacks can mostly be attributed to bots trying their luck on finding devices with port 23 open and then using the default username and password. That or a brute-force attack to gain access into the target network.
Source: Akamai State of the Internet report
While the data in the report reminds the network admin not to leave unused ports open, it also shows that HTTP and HTTPS, both of which are open in most enterprise networks, too are targeted for attack. And then, port 23 or none of the top 10 ports listed might be the ones used to target your network. It can be a different random port which you might have left open inadvertently or had to leave open to facilitate a business service. Of course, it is not possible to block all ingress traffic originating from the WAN to your network.
Firewalls and Intrusion Detection/Prevention Systems (IDS/IPS) enhance your network’s security and are a necessity. But they may not successfully protect your network every time—to name a few, everyone remembers what happened to Sony, Home Depot, and Target! These organizations definitely had security measures in place to protect against malware and other threats—but despite their efforts, the breaches still occurred. This shows that malware and other network threats are getting smarter every day and the traditional methods of security using firewalls and IDS/IPS alone are not sufficient. The work around?
A New Security Layer:
In addition to firewalls and intrusion detection systems, add a 3rd layer of security that can detect threats and attacks that have breached your defense. A layer that looks at the behavior of network traffic to detect anomalies, such as malware, hacking, data theft, and DDoS attacks.
With Network Behavior Anomaly Detection or NBAD, it is possible to detect anomalies that get past the firewall and IDS/IPS systems. NBAD tracks traffic behavior and alerts you if there is unusual or out of the ordinary activity. For example, traffic originating from invalid IP addresses, traffic on one port from one system to many, TCP or UDP packets whose size is less than the least expected value, etc., are all network behavior anomalies. NBAD is further enhanced, when individual systems in the network are monitored for behavior anomalies.
Enterprises can get started with NBAD on their own using traffic flow data, network performance data, and log analysis.
Flow technologies, such as NetFlow, sFlow, J-Flow, or IPFIX carries information about the IP conversations with details like source and destination IP addresses, ports, protocol, volume, number of packets, etc. The data can then be used to track behavior anomalies, such as burst of packets, traffic from invalid IP addresses, malformed packets, etc.
Network performance data can also help discover network anomalies. If there were sudden voice call drops, it could be due to fully utilized links which in turn could possibly be a DDoS attack.
While flow based analysis of traffic is the most widely used method for NBAD, log analysis from various elements in the network including user systems can add value to network behavior analysis. With a log analysis tool that analyzes logs and extrapolates information based on correlation, the admin can pin-point the source of threats within the network and take preventive measures before major damage occurs.
While you are still waiting to find a dedicated NBAD tool that really does what you need, leverage existing technologies and tools for your own network behavior analysis engine. So, what are you starting with? NetFlow or log analysis?